As the rounds go on the players get more tired, and as they get more tired the blunders start to accumulate. There were four decisive games in Khanty-Mansiysk today, and in at least three of them the errors were more unforced than forced.
Fabiano Caruana started the round in clear first and had White against Dmitry Jakovenko, and things were proceeding smoothly. Jakovenko sacrificed a pawn to make things messy, but it really looked like a two-result game: either Caruana would grind out a victory, or Black would draw either by having enough counterplay to keep White from doing what he wanted or by reaching a drawn opposite-colored bishop ending. After a while it looked more like the latter than the former, but a Black win was out of the question until Caruana's 36.Qb3??, overlooking or underestimating 36...c4. (The problem is that 37.dxc4 Qa5 wins a piece.) Black went from being a pawn down to a monster pawn up, and when White decided to avoid the queen ending with 41.Rf2? Black was on his way to delivering mate.
Fortunately for Caruana, he's still in a first place tie. (Jakovenko, surprisingly, is just a half a point behind.) With a win either Leinier Dominguez or Sergey Karjakin could have leapfrogged into first, but neither did. Dominguez was worse forever against Evgeny Tomashevsky, but defended pretty much perfectly and drew in 101 moves. He is thus tied for first. As for Karjakin, he lost to Alexander Grischuk, and in one move. He had come under some pressure near the end of the time control, but after Grischuk's 39.Qe8+ the position would be about equal after 39...Kh6. Instead, Karjakin played 39...Nf7??, still with several minutes on the clock, and resigned after 40.Qg8.
The third member of the leading triumvirate is Hikaru Nakamura. His opponent was Baadur Jobava, so you know it must have been an exciting game. Jobava flung his kingside pawns in the opening, but something went wrong and Nakamura was soon better - much better. He was well on his way to a pretty straightforward victory until he played 44...exf5; either 44...gxf5 or especially 44...dxe4 gxf5 would have made his life much simpler. The point is that Nakamura wound up with a group of pawns around his king that constituted a sort of do-it-yourself mating net, and while there were other improvements available to Nakamura later on Jobava had loads of counterplay based on Black's terrible king.
A key moment came on move 67, when Jobava played 67.Ke7. Given his intention to play Nxg6 next, he should have played 67.Kd5 instead, when the same continuation as in the game would lead to a draw: 67.Kd5 Kg8 68.Nxg6 Rxg6 69.Rxh5 Rg4 70.Ke6! g6 71.Rh1! Kg7 72.Ke5! Ra4 73.Rh2 etc., and when Black plays ...g5 White plays Kf5, and then it's trivial. In the 67.Ke7(?) version, the sac failed, as there was no way to get the king back or to create a sort of mutual standoff where Black must let the king back in order to make progress. Maybe White could have held if he hadn't played 68.Nxg6 - I'll leave that to you guys to work out.
The final winner of the day was Boris Gelfand, whose win over Peter Svidler was his first win of the tournament - and despite that he's just half a point out of first. This was one of those games where the evaluation moved in waves: equal at the start of the game, then White (Gelfand) was much better (maybe winning), then Black got back to equal, then White was better again, then equal, then White was better...and the third time, Svidler couldn't get back on the wave and he - or rather, his position - went under.
The last game of the day was an uneventful draw between Maxime Vachier-Lagrave and Anish Giri. For Vachier-Lagrave, it was enough to break a four-game losing streak, while Giri was probably tired from yesterday's marathon with Tomashevsky and reasonably happy to get an easy draw with Black going into the rest day. When the players resume battle on Sunday, these will be the pairings: